Rorty I 255
FodorVsRyle/Rorty: a simple story about learned associations will not be enough: the expectancy system would have to be abstract and complicated in the same sense. Because the recognized identities are surprisingly independent of the physical uniformities of stimuli among one another.
>
Stimuli.
Rorty I 255
RortyVsFodor: suppose we needed an abstract recipe for the recognition of similarities among potentially infinite differences. Why must the recipe be so abstract at all? Presumably, we must be able to identify similarities. But then we do not need the idea of a "non-abstract" recipe, because every recipe must be able to do this.
E.g. Rorty: the possible qualitative differences of the contents of a package of chocolate cookies are also potentially infinite.
Rorty: so if we talk about "complex expectation systems" or programs or control systems at all, we are always talking about something abstract.
Dilemma: either the explanation of the acquisition of such control systems requires the postulation of other control systems or they are not learned at all.
>
Learning.
Either:
1) There is infinite regress, because what is true for recognition would also need to apply for learning.
Or 2) we end up back at Ryle: people have an ability that they have not learned.
Rorty I 269
Fodor rehabilitates the traditional British theories of perception: "it is an empirical question whether psychological processes are computer processes! If they are, our perception must work in a way that a description of the environment that is not done in a vocabulary whose terms represent values of physical variables, is calculated on the basis of a description made in such a vocabulary.
>
Perception, >
Vocabulary.
I Rorty 269
Fodor: why should there not be stimuli for the whole organism? Then you could discover a stimulus variant "bottle".
Perception requires the choice of an independent vocabulary for the representation of the inputs. Fodor's thesis: all perceptual knowledge is transferred by the activity of sensory transducers.
Rorty I 269
Fodor: if we do not want to realize the talk of the information processing, we need to use something that our subject need not necessarily know as its input. Rorty question: could it turn out that the input is not completed on the retina but half way or elsewhere? Fodor presumes: yes, it just depends on which design of the black box the organism can best be considered to be split into converter and processor, so that the best theory comes out.
